0xDECAFBAD - Tag: yahoo

Delicious 2.0 is more than a pretty new face

2008-08-04T17:21:55+00:00

I'm no longer at Yahoo! and I no longer work on Delicious, but I'm still a huge supporter. And, since I'm pretty sure everyone over there is either burnt out or still insanely busy at the moment, it might be awhile before anyone tells the full story of what this relaunch offers. As it is, even I only know a bit of what all went into this effort.

So, off the top of my head, I thought I might point out just a few of the easily-missed improvements the new site offers beyond the great new love-it-or-hate-it visual redesign that seems to have occupied most of the discussion I've seen so far:

Search works and is incredibly fast. In fact, search may really be the real star of the redesign show here—especially since del.icio.us often took 30 seconds or more for a simple search, rendering it all but useless. Today, though, it's at ludicrous speed in comparison—and so finally, the real power of search applied to social bookmarking might start to shine with the critical mass of content found by real people using Delicious.

To scratch my own itch, I've created an unofficial OpenSearch search engine plugin for Delicious on Mycroft. Though I think it comes along with the browser extensions, I've yet to find this for autodiscovery from the site itself.

The notes field in bookmarks has been expanded from 255 to 1000 characters. Yes, this means that you can now include long running quotes from pages, or complete paragraphs of rambling discourse.

The feeds have all been overhauled and reorganized. An attempt at backward compatibility was made, but the old feed URLs are all deprecated. Replacing these, there's now a common and consistent URL namespace for feeds across formats.

Almost all RSS feeds have JSONP counterparts, and further feed formats could be considered. Additionally, the old mix of RSS 1.0 and 2.0 has been dropped in favor of RSS 2.0 format across the board to support podcast and media enclosure elements consistently. The linkrolls, tagrolls, network badges, and tagometers now all use the new JSONP feeds—and the widgets can be examined as example code in using the feeds.

Tag bundles can now be viewed as combined bookmark views, complete with feeds. This augments bundles from a simple visual organization tool to a more powerful content aggregation function. Personally, I never had much use for tag bundles until now, but since they can actually be used to partition tags and bookmarks I might actually take the time to use them. Check out these example URLs:

Bundles now work for tags, network contacts, and subscriptions. This means, for example, that you can partition your network contacts into topical groups. From there, you can subscribe to those partitioned bookmarks in different folders in readers, or use the bundled bookmark views in mashups through JSON feeds. Check out these example URLs:

Network member and fan feeds now include the date when the contact was added. Pro tip: Subscribe to fan feeds to see when new people have started following your bookmarks.

Several previously undocumented parts of the V1 API have now been documented. In particular, the following new parameter combinations have been used with the browser extensions for primitive bookmark sync:

https://api.del.icio.us/v1/posts/all?hashes
https://api.del.icio.us/v1/posts/all?meta=yes
https://api.del.icio.us/v1/posts/all?fromdt={DT}&todt={DT}
https://api.del.icio.us/v1/posts/all?start={##}&results={##}

The posts/all API URL works again for users with large collections. With my 11k+ bookmarks, del.icio.us was keeling over with the attempt to assemble and deliver my entire collection with a posts/all call. Now, however, Delicious 2.0 appears able to handle this call for my account.

Easter eggs have been rotated and recalibrated. No, I'm not going to tell you what or where or how many they are.

Archived Comments

Aristotle Pagaltzis
2008-08-05T02:31:52

Since I’m pretty sure everyone over there is either burnt out or still insanely busy at the moment, it might be awhile before anyone tells the full story of what this relaunch offers.

Admit it, Les, you just got impatient after building a bunch of shiny toys that no one at Yahoo! seemed to be coming forth to tell the world about them. :) (I would be miffed too.)
Yining
2008-08-05T06:51:44

one thing in the new version that disappoints me is the tab-to-complete feature is gone. It was replaced with a drop-down list which not only costs me more finger/hand movement but also breaks my daily usage habit of del.icio.us. Oh, maybe I can write a GMscript to scratch my own itch....

Btw, regarding the search, I find using a keyworded bookmark is much faster. For example, I could setup keyword 'del' and just type 'del webdev+openid' to search all my bookmarks tagged with these two keywords. One plus point, it works on SeaMonkey too (which doesn't support opensearch yet).
Nick
2008-08-05T06:56:05

Yay! So glad you wrote this, Les!
Paul D. Waite
2008-08-05T17:07:03

one thing in the new version that disappoints me is the tab-to-complete feature is gone. It was replaced with a drop-down list which not only costs me more finger/hand movement but also breaks my daily usage habit of del.icio.us

Amen.

Also,the option to delete a bookmark from its own page seems to have disappeared. I have to find it in my list of bookmarks, which isn’t always easy when it’s been there for a couple of weeks.
Paul D. Waite
2008-08-05T17:09:25

Sorry, that sounded a bit negative. I do love how autocomplete now only offers me tags that I’ve used. And it is kinda neat to see how many times I’ve used each one, but I’d give that up to have the old tab-to-autocomplete back. Way too much thinking involved with the menu.
Niv Singer
2008-08-05T18:52:38

I couldn't find the "Network member and fan feeds" - where are they??
l.m.orchard
2008-08-05T19:03:00

@Niv: Not all feeds are linked in from pages - check out the feeds help page on delicious.com. For example:

http://feeds.delicious.com/v2/rss/networkmembers/nivsin

http://feeds.delicious.com/v2/rss/networkfans/nivsin
Philip Dorrell
2008-08-06T02:01:52

It seems that the posting API's have all stopped accepting a "tags" parameter, which partly breaks my Firefox "deli" search keyword.

e.g. "deli toread math" used to fill in tags "toread" and "math", now it doesn't.

Also, does delicious have any equivalent to Google's site: prefix? That would be really good for a "most popular pages on my website" option.
l.m.orchard
2008-08-06T21:41:06

@philip Yup, that's broken for me too. See also: http://decafbad.com/blog/2008/08/02/delicious-20-legacy-bookmarklet-fix

As for something like site:, I think host: used to work but it seems no longer with the new search implementation.
Jon Williams
2008-08-22T18:38:43

Network member and fan feeds now include the date when the contact was added. Pro tip: Subscribe to fan feeds to see when new people have started following your bookmarks.

Where is this? (I feel silly not being able to find fan feeds)

suggestion: tinfoil hat for digg mind control laser defense

2007-02-15T16:34:27+00:00

We call it a Suggestion Board — you can browse suggestions from other site visitors or post your own. Digg-style voting means we can quickly discover what’s most important to users.

Source: It takes two to Tango - Yodel Anecdotal

Sweet merciful crap are Digg fans a scary, groupthink-y bunch. I work at Yahoo! — so I guess I'm vulnerable to bias — but it seems like the guys on this team just can't get a break. They open up public suggestion boards to get open feedback on products — something that I think is very cool and clueful. They took inspiration from Digg and other prior examples of web-based voting, and what do they get for their trouble? A few hundred lemmings marching to the beat of "OMFG HAX YOU STOLE DIGGS SITE!"

It's not a major news site; it's not like the front page of Yahoo! turned into a Digg clone and they started offering to pay top Digg users. Seriously - it's just a suggestion board with simple flat comment threads that invites voting to surface good suggestions. So it's got a big thumbs-up button reminiscent of Digg's counter badges - they mentioned that it's like Digg. That's inspiration, maybe flattery-level imitation. But theft? If I were a thief, would I tell you where I stole my stuff from?

Archived Comments

Brett
2007-02-15T17:06:18

Digg always makes me think of: http://sethgodin.typepad.com/seths_blog/2005/01/mob_justice.html

"piling on is safe and fun"
Craig
2007-02-15T18:37:08

Delete this if you think this is off topic, but am I the only person on the planet that doesn't quite understand the allure of Digg? Call me stodgy, but I kinda like the editor-in-control feel of Slashdot vs the "OMFG! KITTIES!" attention-cluster-bomb nature of Digg. I've subscribed several times via RSS, and haven't remained subscribed more than a day or two because of the complete crap-flood in my RSS reader.
Pete Prodoehl
2007-02-15T21:39:24

Don't tell the Digg mobb about Jyte, I mean, besides voting buttons it even has a 4 letter name like Digg does! OMG C0P13RS!
Aristotle Pagaltzis
2007-02-16T00:14:26

I assume you saw Digg scares me (403 Go Away!)?

Thoughts on Pipes on the Web - Part II

2007-02-15T08:49:56+00:00

In the last post, I expressed concern for my own well-being over a lack of head-over-heels love for Yahoo! Pipes. On the surface, I think it's because it's not a freshly discovered concept for me, and also probably because I'm tired and a bit hacked out right now. But, I think there's a bit more to it.

First of all, I'm not impressed by GUI builders for most things. Don't get me wrong: That Pipes GUI is pretty sweet and quite an impressive implementation — but as a rule, such things don't quite get my geek on. So, there's that.

Another thing is that I've done something kinda like Yahoo! Pipes, called FeedMagick — only, it doesn't have a GUI and I mostly abandoned it after releasing the code and having used it for a project at my old job. It's one of those serial enthusiasms that I've figured I'd circle back to eventually. I still think it's a pretty cool idea.

All kudos to the Yahoo! Pipes team though: Unlike me, they've actually got a living and breathing project — which trumps a paged-to-disk serial enthusiasm most days.

So what else is really curbing my enthusiasm? Well, Yahoo! Pipes looks like a pretty self-contained pipes engine — data goes in one end, frobnication happens in the middle, data comes out the other side.

But, what I like about the notion of the URL-line is that you can take one URL and supply it as a parameter to another URL — making messy pipelined URLs while building a crazy web-wide distributed execution environment powered by HTTP and REST. This is the kernel of the notion that I think really excites me about pipe on the web — I just haven't had a chance to do much with it lately.

Depending on the perspective, true pipes on the web — that is, URLs fetching URLs — look to me like functional programming ala Lisp. Consider an idempotent GET request as a pure function call with no side-effects. Then, consider a GET request that accepts a URL as a parameter — it's a nested function call: the outer GET must make an inner GET to fetch the parameter-supplied URL. Give the inner URL another URL as a parameter, and you've got yet another nested function call.

But, if you like, ignore the theoretical benefits of functional programming — flip the nested function calls inside out and you've got a pipe. And, since you're using HTTP GET, you can get all the benefits of HTTP — like caching of execution results and a web full of distributed processing nodes, among other things.

I haven't explored Yahoo! Pipes deeply enough yet, so maybe I'm missing the features that pipe authors can used to call on other distributed pipe elements out on the web at large. But, I think that's what ultimately gets me psyched about pipes the web and hasn't yet for Yahoo! Pipes.

And, of course, I think I'm just a little hacked out and tired of whizbang new stuff right now. :)

Archived Comments

John Evans
2007-02-15T13:59:01

The thing that struck me about Pipes is that it seemed rather limiting to someone who can code but too geeky to someone who can't. Maybe I am wrong but I wonder that once the hype dies down who exactly the audience will be long term.
Stuart Robinson
2007-02-15T19:35:41

Don't be so apologetic! Conceptually it's cool, but the UI quickly becomes frustrating for anyone with text-based coding experience. I guess this is the web's visual basic and we are all going to be more interested in it's perl.

Thoughts on Pipes on the Web

2007-02-15T08:09:56+00:00

By way of my meme tracker, I found something like 15 entries linking to Yahoo! Pipes from my feed subscriptions. That's crazy — I've usually never seen more than 3-5 feeds linking to anyone thing, maybe as many as 8 if it's something really hot. I read about it over morning coffee and aggregator skimming; I heard about it on TWiT on the bus to work; I was told all about it in the hallway once I got to work. Yahoo! Pipes has been hard to miss.

With that in mind, it seems weird to me that I didn't get as excited about the release of Pipes as everyone else in my local blogosphere seems to have. Pipes has all the ingredients of something that should really set me off. Pipes, filters, mashups, RSS, Atom, RDF — (singing) these are a few of my favorite things! So, what gives? I'm genuinely worried that I'm sick, depressed, losing it, or otherwise replaced by a pod alien. Possibly all the above.

I knew about Pipes as an internal project at Yahoo! quite awhile ago — possibly even in my first week or two after having joined. I'm fuzzy on the timing here, but it's not important. Someone mentioned it to me at some point, I poked at it a bit from the internal network, and thought it was neat / keen. My main impression at the time was, "Cool, someone's doing FeedMagick, but done right and with a slick visual layout tool."

Then, I got really busy with del.icio.us and wandered off. So, the release wasn't big surprising news to me — my reaction was something like, "Huh, so they did a public release of that? Nice." Clearly I need more caffeine.

The thing is that the idea of pipes on the web is not itself a new concept to me. I think my XmlRpcFilteringPipes seems kinda silly in retrospect, and I'm all about the Power of the URL-line nowadays — but suffice it to say that I've been thinking about pipes on the web since as long ago as 2002 or so.

So, while I'm all jaded and apparently incapable of feeling the wonder, this seems to be one of the biggest revelations to people playing with Yahoo! Pipes:

Pipes? On the web? BRILLIANT!

So, it's not you, it's me. I think I've got a serious deficit of enthusiasm for what's really a pretty cool thing. I want to think a bit more about why that is.

Archived Comments

Kenneth
2007-02-15T14:06:37

http://yakubovich.blogspot.com/2007/01/visipipe-visual-version-of-core-utils.html

Might have been buried in all the buzz about Yahoo, but this sounds like a neat idea.

Kenneth Writer
Jeff Daly
2007-02-16T01:08:54

You're not alone. I was excited when I learned about Pipes, but ultimately I was let down by the lack of parsing tools. Pipes is seriously limited without any XPath/regular expression parsing capabilities. Fortunately, there is a post on XPath in the feature request page.

yahoo TV refreshed, redux and rejected

2006-12-02T20:25:28+00:00

Okay, I'm sick of it now, like many others. I was dazzled by the AJAX pagination technique at first, because I think it's an interesting advance in AJAX in general. But, I've been a user of the Yahoo! TV Grid - er Listings - for years now. It's useless to me now, even after I re-discovered its location after my bookmark broke. I hate to be harsh, but I'd outsourced a part of my brain the Yahoo! TV Grid, and suddenly that part of my brain is damaged.

Update: I've just noticed that this entry's linked to by Dave Winer, so maybe I'm in danger of being read. :) To be not-so-harsh, I had posted some suggestions to the above-linked blog entry, but I'm not sure if they'll get through the moderation queue. Nothing sinister expected or asserted, just an overload of voices to which I'm but a late piler-on.

Basically, I'd be made happy by Yahoo! TV if they treated the TV listings grid as a valued entry point with in-context links to discussions and reviews from the the shows listed, and if they dropped the AJAX pagination to present the entire grid at once for in-browser search and random access visual scanning. This relatively new AJAX pagination technique is pretty sweet, but better applied to data sets ordered by relevance and / or expected to be traversed in one direction.

Also, a note in general to other redesign teams: Watch your entry points and don't break bookmarks unexpectedly.

Archived Comments

Jeremy Zawodny
2006-12-03T04:09:32

Your "like many others" link is 404ing.
l.m.orchard
2006-12-03T05:48:00

Doh't. Used some parens when I mean square braces in my Markdown.

yahoo TV refreshed

2006-11-30T23:02:20+00:00

I'm a little annoyed that the Yahoo! TV refresh last night broke my bookmarks. We're luddites who've rejected Digital Cable and On-Screen Guides for the last few years - and Yahoo! TV has been our solution for that. But, oh well. I've got a new TV guide bookmark now, and all better.

On the other hand, I'm a bit impressed and entertained that the TV guide grid uses an AJAX-ian pagination approach like this one.

Of course, on the other other hand, it's pretty slow now. I found that if I scroll down the the bottom of the page, really really fast, everything gradually loads up. That seems somewhat suboptimal.

Archived Comments

Michal Migurski
2006-12-01T01:00:30

I'm surprised that the new site breaks old links - seems like having Del.icio.us within the company should underscore the importance of permanent URLs.

Yahoo for stalking iPods in the hands of couriers

2005-11-03T18:29:12+00:00

If the new Yahoo! Maps were really cool, they'd've already done a mashup with DHL to let me obsessively stalk the courier in whose truck my returning iPod repair currently rests.

Come on, get with the program! When we've got GPS watches, "Status: With delivery courier / Location: (blank)" is no way to track a package! In the meantime, my repetitive information injury is flaring up as I refresh, refresh, refresh, refresh...